i'm considering this the first candidate for a 1.0 release. as usual,
i'm welcoming bug reports and patches which fix problems.
$Id: CHANGES 1.45 Sun, 19 Aug 2001 15:19:05 -0400 trockij $
Changes between mon-0.38.21 and mon-0.99.1
Sun Aug 19 15:18:55 EDT 2001
--------------------------------------------
******DEFAULT BEHAVIOR HAS CHANGED FOR THE FOLLOWING FEATURES************
the following two defaults were changed, since they seem to be unintuitive
to most people, based on feedback given on the mailing list.
-the old "comp_alerts" is now the default. to get the old
behavior, specify "no_comp_alerts" in the period section.
-the default is now the old "summary" behavior for alertafter. that
means that for successive failures with "alertafter" used to suppress
multiple alerts, only the summary line will be used to short-circuit
the alert suppression. to get the old behavior, append "no_summary"
to the alertafter line. the old "summary" syntax is still permitted
to help w/backwards compatibility.
*************************************************************************
-cleaned up config parsing a bit
-updates to up_rtt.monitor, added traceroute.monitor, smtp3.monitor,
http_tpp.monitor, file_change.monitor
-fixed problem where upalerts were not sent for ack'd failures
-updated the sample etc/auth.cf to give examples of trap authentication
-updated man page for mon to include better explanation of
auth.cf syntax.
-formatting updates to monshow, added "summary-len" option, html
fixes
-fixed problem where server responded twice with an extra "220 ok"
after doing a reload
-rewrote fping.monitor to return more verbose output, and to sort
the failed hosts on the summary line. this was wreaking havoc
with "alertevery", since the order of the failed hosts in the
summary might change, even though the same hosts were failing on
successive tests. added "-s ms" option which will consider hosts
with a response time greater than "ms" milliseconds as failures.
added "-a" option to fail only when all hosts fail, and "-T" to call
traceroute on each failed host. "-h" lists options.
-made nearly all monitors output their summary line (if it is a list
of hosts) in sorted order.
-updated man page for mon with more detail on the behavior of "alertevery"
and "alertevery ... summary"
-added xedia-ipsec-tunnel.monitor to monitor site-to-site ipsec tunnels
on a Xedia AP450 router.
-silkworm.monitor recognizes different brocade OEM'd fcal switches,
ignores "absent" sensors, and has a work-around for the braindead
behavior of swFCPortAdmStatus to detect offline ports.
-fix to msql-mysql.monitor to allow --port to override default port.
submitted by Adrian Phillips <[EMAIL PROTECTED]>
-stdout and stderr now can be sent to a file by specifying a filename
in the variable "monerrfile".
submitted by Ed Ravin <[EMAIL PROTECTED]>
-updated dns.monitor to output only the failed hosts on the
summary line.
"test config" fix,
new authentication directives "!" and "AUTH_ANY". "AUTH_ANY",
check and warnings for hostgroups which are defined but never used
more descriptive error when m4 is not found
removed second definition of disen_host and load_stat
"alertafter timeval" patch, alerts for period will only be called if
the service has been in a failure state for more than the length
of time desribed by the interval, regardless of the number of failures
noticed within that interval.
submitted by Andrew Ryan <[EMAIL PROTECTED]>
-more verbose error when bind(2) failure
tyop fixes to mon.1
updated COPYRIGHT
mon.1 is now mon.8, and references to mon.1 changed accordingly
update to mon.d/Makefile to use $CFLAGS and $LDFLAGS
silence some warnings in rpc.monitor.c
add /usr/local/lib to standard search paths for alert.d and mon.d,
and updated mon.8
make monshow run under taint mode, fixes view directory to match the docs
default server for moncmd and monshow is now localhost
http.monitor accepts a 302 status (moved temporarily)
fixed --auth in monshow
reboot.monitor now uses $MON_STATEDIR as the default state directory,
and "reboot.monitor" (not "state") as default state file.
FD_CLOEXEC fix
update to monshow.1
submitted by Roderick Schertler <[EMAIL PROTECTED]>
-fix to pop3.monitor to produce more verbose errors
fix to reboot.monitor to add --verbose option
submitted by Ed Ravin <[EMAIL PROTECTED]>
-qpage.alert accepts "-v" option for verbose
smtp.monitor has increased verbosity of failure details
submitted by Steve Siirila <[EMAIL PROTECTED]>
-re-wrote Steve Siirila's mon.monitor to use Mon::Client and put it
in mon.d
-patch to do proper syslog handling on openbsd,
MON_DEPEND_STATUS env variable passed to monitors
submitted by Mark D. Nagel <[EMAIL PROTECTED]>
-added "failure_interval" functionality. i actually re-wrote
the patch to make it a bit more proper, and renamed the
parameter from "alertintervalcheck" to "failure_interval"
for clarity.
submitted by CHASSERIAU JeanLuc <[EMAIL PROTECTED]>
-netappfree.monitor changes
Allows the monitor to give more verbose error messages which
will handle multiple volumes. Instead of reporting:
"1.0GB free on <filer>"
it will now say:
"1.0GB free on <filer>:/vol/<volume>"
Fixes a bug where multiple alerts from a single filer would cause
multiple entries in the summary line. Allows the monitor to handle
the case where the NetApp MIB isn't available to the script.
added na_quota.monitor. trockij made some small changes to it
so that it will allow disable and enable to work.
submitted by Theo Van Dinter <[EMAIL PROTECTED]>
